Temperature Sensing to Switch on/off an Induction Heater

[Kaleb Jackson]

Hello, I'm having trouble thinking of ways that I can measure the temperature of an object that I am heating using an Induction Heater. A little background: The coil and workpiece will be submerged in liquid, and due to most thermometers being metal it won't work with the induction heater. I'm basically just trying to make a temperature feedback circuit so I can pretty much make my induction heater auto hover around a certain range of temps but am having trouble figuring out ways to determine temperature of the object being heated when submerged into liquid (an acidic solution to be more specific). Any help, ideas, or interesting facts pertaining to this would be awesome and super appreciated. Hopefully anorlunda or whoever the mod is approves this thread <3

 

[.Scott]

This is what I would try:
Rather than measure the temperature of the work piece directly, attempt to estimate its temperature by how much it is heating the liquid.
So keep the liquid circulating in a predictable way that tents to even out its temperature. You work piece will be dispensing heat into the body of liquid and the liquid will be conducting it away.

The temperature of the liquid will tend to be at some ratio of the ambient temperature and the work piece temperature.
It will also exhibit a lag in temperature compared to the work piece.

So you need to characterize this behavior through experimentation. For example, let's say your target
work piece temperature is 500 and ambient is 50. You may discover that if you keep the work piece at the target temperature, the temperature in the circulating liquid hangs out at 440.
In the long run, your liquid set point will be 440. But when heat is first applied, that set point will be lower, because the liquid has not had time to warm up. So you need to guard against overheating - by properly modelling the ideal liquid temperature profile during the heating process.

 

[Kaleb Jackson]

This is what I would try:
Rather than measure the temperature of the work piece directly, attempt to estimate its temperature by how much it is heating the liquid.
So keep the liquid circulating in a predictable way that tents to even out its temperature. You work piece will be dispensing heat into the body of liquid and the liquid will be conducting it away.

The temperature of the liquid will tend to be at some ratio of the ambient temperature and the work piece temperature.
It will also exhibit a lag in temperature compared to the work piece.

So you need to characterize this behavior through experimentation. For example, let's say your target
work piece temperature is 500 and ambient is 50. You may discover that if you keep the work piece at the target temperature, the temperature in the circulating liquid hangs out at 440.
In the long run, your liquid set point will be 440. But when heat is first applied, that set point will be lower, because the liquid has not had time to warm up. So you need to guard against overheating - by properly modelling the ideal liquid temperature profile during the heating process.

That was one of the only ideas that I felt could really hold true over time, it is a bit of lagged feedback like you said but it's better than nothing and can definitely be used at least as a back up. If I go with this method then I'll probably use 2 separate modes of 'on', one being a startup mode and the other being an 'equilibrium' mode where it's met temps and just trying to maintain. Does this sound good in your opinion? Any modifications you would make?

 

[.Scott]

If I go with this method then I'll probably use 2 separate modes of 'on', one being a startup mode and the other being an 'equilibrium' mode where it's met temps and just trying to maintain. Does this sound good in your opinion? Any modifications you would make?

Two modes may be fine. In the start mode part 1, you would apply high current for a period of time proportional to the mass of the work piece. Then, start mode part 2, holds a lower current that is predicted to be good for maintenance. Then switch to a maintenance mode with feedback that slowly adjusts temperature in reaction to changes in the liquid temperature. But in the startup modes, the response to a high liquid temperature should be to shut things down so the operator can identify what is going wrong.

If you can put this under computer control, you can do something more precise.

 

[Kaleb Jackson]

Two modes may be fine. In the start mode part 1, you would apply high current for a period of time proportional to the mass of the work piece. Then, start mode part 2, holds a lower current that is predicted to be good for maintenance. Then switch to a maintenance mode with feedback that slowly adjusts temperature in reaction to changes in the liquid temperature. But in the startup modes, the response to a high liquid temperature should be to shut things down so the operator can identify what is going wrong.

If you can put this under computer control, you can do something more precise.

Good ideas here, Scott. I'll take all this into consideration and let you know how it goes if I am able to get something working in the next couple weeks. Thanks for the well thought out response!

 

[Tom.G]

A thermocouple on/in the object could also work if you bring the leads axially out of the induction coil. Their output is DC so any residual interference is easily filtered out of the signal.

Probably the easiest way to find a compatible one is to ask manufacturers what they have to meet your requirements.

Thermocouples are available in a range of materials and you will have to research which ones, if any, can withstand your particular acid. They are also available with various electrical insulation on the leads, such as Fiberglass.

 

[Tom.G]

Or put a surface mount thermostat on the outside of the container.
Disadvantage is a longer time lag.
Advantage fewer $$$.

 

[Rive]

What is the target temperature range?
The trick what is used in the (old style) fridge thermostats might be some use for you. With a pressure sensor instead of contacts.

 

[Baluncore]

... but am having trouble figuring out ways to determine temperature of the object being heated when submerged into liquid (an acidic solution to be more specific) ...

An acid solution is also conductive, so you will also be heating the solution with the induction coil.
If the metal temperature exceeds the boiling point of the liquid, bubbles will form that will change the heating pattern.
Can you change the solution to adjust the BP so as to regulate the temperature?
What temperature do you need to maintain?
What is the boiling point of the solution?
How will you manage the acidic fumes?
Induction heating may not be applicable in your situation.

 

[Kaleb Jackson]

Yeah I'm thinking Induction heating is hardly the most applicable fit to the situation, but my employer wants to try it so I'm doing what I can with as few resources as possible as to not waste time AND money. The acidic solution is giving us the biggest issue, and with the temp we're trying to get to it's going to make everything pretty difficult. Here's some explanations that will hopefully help:
- The solution is in a big sink essentially with our workpieces in the solution
- We cannot change the solution, it's pretty specific to what the company does and really can't be altered (to my knowledge)
- We need the temp to hover around boiling for water so about 212 F (again with the bubbles issue)
- I'm not sure of the boiling point of the solution, we haven't been able to run real world tests yet but will get that info ASAP
- There's a lid on the 'sink', and we were hoping to coat or surround the coils with a heat resistant plastic in order to get right next to the workpieces (as opposed to outside the sink) without getting acid straight up on the coil or circuit
Issues:
- Fumes could easily destroy the circuit, or coil at least
- Magnetic Field dissipates extremely quickly/power drop off is fast
- Portion of the workpiece to be heated is fairly specific, though we can try different coil designs
- No temp control whatsoever (other than light switch buffer between the PSU and the heating circuit)

Thanks for the help, man. I really appreciate the ideas!

 

[Kaleb Jackson]

A thermocouple on/in the object could also work if you bring the leads axially out of the induction coil. Their output is DC so any residual interference is easily filtered out of the signal.

Probably the easiest way to find a compatible one is to ask manufacturers what they have to meet your requirements.

Thermocouples are available in a range of materials and you will have to research which ones, if any, can withstand your particular acid. They are also available with various electrical insulation on the leads, such as Fiberglass.

Not a bad idea, I'll look at some thermocouples and see if there are any that would seem fit for the solution. Sadly I only know the types of acid they use since the ratios of acid to whatever else they put in is basically proprietary and I'm just a seasonal worker.

 

[Kaleb Jackson]

What is the target temperature range?
The trick what is used in the (old style) fridge thermostats might be some use for you. With a pressure sensor instead of contacts.

I'm looking at roughly 212 F (boiling point for water) for the temp range. You think it would work in a mostly non-pressurized container? I'm a bit unfamiliar with the fridge thermostats (but I'll go research them).